Three different types of chirality-driven crystallization within the series of uniformly substituted phenyl glycerol ethers

Article abstract of DOI:10.1002/chir.20648

Seven chiral arylglycerol ethers 2-R-C₆H₄-O-CH₂CH(OH)CH₂OH (R 5 H, Me, Et, Allyl, n-Pr, i-Pr, tert-Bu) were synthesized in racemic and scalemic form. The IR spectra, melting points, and enthalpies of fusion for racemic and scalemic samples of every species were measured, the entropies of enantiomers mixing in the liquid state and Gibbs free energies of a racemic compound formation were derived and binary phase diagrams were reconstructed for the whole family. Solid racemic compounds stabilities were ranked for the four substances. Spontaneous resolution was established for the registered chiral drug mephenesin and its ethyl analogue. Metastable anomalous conglomerate, forming crystals having three independent R and one independent S molecules in the unit cell, is formed during solution crystallization of tert-butyl derivative; metastable phase transforms slowly into traditional racemic conglomerate. Chirality 20:1092-1103, 2008.

Full text of DOI:10.1002/chir.20648

CHIRALITY 20:1092–1103 (2008)
Three Different Types of Chirality-Driven Crystallization Within
the Series of Uniformly Substituted Phenyl Glycerol Ethers
*
ALEXANDER A. BREDIKHIN, ZEMFIRA A. BREDIKHINA, VICTORINA G. NOVIKOVA, ALEXANDER V. PASHAGIN,
DMITRY V. ZAKHARYCHEV, AND AIDAR T. GUBAIDULLIN
A.E. Arbuzov Institute of Organic and Physical Chemistry, Russian Academy of Sciences, Kazan, Russian Federation
Dedicated to the 160th Anniversary of Louis Pasteur’s Discovery
ABSTRACT
Seven chiral arylglycerol ethers 2-R-C6H4-O-CH2CH(OH)CH2OH (R 5
H, Me, Et, Allyl, n-Pr, i-Pr, tert-Bu) were synthesized in racemic and scalemic form. The
IR spectra, melting points, and enthalpies of fusion for racemic and scalemic samples of
every species were measured, the entropies of enantiomers mixing in the liquid state
and Gibbs free energies of a racemic compound formation were derived and binary
phase diagrams were reconstructed for the whole family. Solid racemic compounds
stabilities were ranked for the four substances. Spontaneous resolution was established
for the registered chiral drug mephenesin and its ethyl analogue. Metastable anomalous
conglomerate, forming crystals having three independent R* and one independent S*
molecules in the unit cell, is formed during solution crystallization of tert-butyl derivative;
metastable phase transforms slowly into traditional racemic conglomerate. Chirality
C
VV
20:1092–1103, 2008.
2008 Wiley-Liss, Inc.
KEY WORDS: spontaneous resolution; anomalous conglomerate; thermochemistry;
X-ray diffraction; preferential crystallization; mephenesin
INTRODUCTION
enantiopure components, and the values of entropy of mix-
ing of enantiomers in the liquid state, DSm1 . The Gibbs free
energy of formation is always negative for a racemic com-
pound, if it can exist, while for a racemic conglomerate
this value must be (but not always is) close to zero. So
obtained the entropy of mixing has no clear interpretation,
but for ideal conglomerate this value must be equal to
Rln2 or 5.75 J K21 mol21. Both of these thermodynamic
characteristics could not be measured directly, but could
be calculated on the basis of the chiral substances melting
point temperatures and fusion enthalpies Tf and DHf.3,5
The method of choice for experimental Tf and DHf deter-
mination is differential scanning calorimetry (DSC).6
Louis Pasteur’s observation of spontaneous resolution of
the tartaric acid particular salt followed by triage separa-
tion of its enantiomers, in his great 1848 experiment1, laid
the groundwork for the modern science of stereochemis-
try. Up to now the nature of spontaneous resolution
remains something of an enigma, and according to Lluꢀısa
Perez-Garcia and David Amabilino ‘‘the understanding and
prediction of spontaneous resolution . . . remains one of
the true challenges for science in the 21st century’’.2 Not
only has an unquenchable thirst for knowledge made
spontaneous resolution so attractive, the practical reasons
are every bit as important as the abstract ones. Based on
this phenomenon, direct methods for production of single
enantiomers, as for instance resolution by entrainment,3,4
refer to natural limits of resolution effectiveness because
of no need of chiral auxiliaries and specialized equipment.
There are different ways to study spontaneous resolu-
tion, tracking the little chemical structure variations
accompanying the crystallization in the series of closely
related chiral compounds is one possibility. The members
of such a series must be selected in such a way that every
compound would have minimal but regular distinctions
from each other. In addition, it would be desirable to have
not only qualitative categories like the ‘‘conglomerate’’ or
‘‘racemic compound’’ for characterization of crystalline
type, but to introduce a quantitative measure allowing us
to rank the observed properties. For this measure one can
use the Gibbs free energy changes, DG0, accompanied by
As regards the choice of the chemical objects, our
recent interests deal with the terminal aryl glycerol ethers
ArOꢀꢀCH2CH(OH)CH2ꢀꢀOH. The reasons must be
pointed out. Firstly, the glycerol ethers and esters are
rather common in the family of lipids. Lipids (fats, plasma-
logenes, membrane forming glycerolipids, and the like)
form the third class, after proteins and carbohydrates, of
‘‘life molecules,’’ hence a spontaneous resolution among
the lipid-like compounds could be related to the problem
of origin of life homochirality.7 Secondly, many physiologi-
Contract grant sponsor: Russian Fund of Basic Research; Contract grant
numbers: 06-03-32508, 07-03-12070
*Correspondence to: A. A. Bredikhin, A.E. Arbuzov Institute of Organic
and Physical Chemistry, Arbuzov st., 8, Kazan, 420088, Russia.
E-mail: baa@iopc.knc.ru
Received for publication 7 March 2008; Accepted 20 June 2008
DOI: 10.1002/chir.20648
Published online 2 September 2008 in Wiley InterScience
the reaction of the racemic compound formation from the (www.interscience.wiley.com).
C
VV
2008 Wiley-Liss, Inc.

THREE TYPES OF CHIRALITY-DRIVEN CRYSTALLIZATION
1093
specific rotation is given in deg mL g21 dm21, and the
concentration of solutions c appears in g (100 ml)21
.
Enantiomeric purity was checked by HPLC analysis per-
formed on a Shimadzu LC-20AD system controller, and
UV monitor 275 nm was used as a detector. The column
used, from Daicel, Inc., was Chiralcel AD-RH (0.46 cm 3
15 cm).
All thermal measurements were performed on a Perkin-
Elmer Diamond DSC model in aluminum pans using a
heating rate of 10 K min21. Mass of the samples amounted
to ꢁ2.5 mg. Temperature scale and heat flux were
calibrated against the data for indium, phenol, and
naphthalene.
Racemic epichlorohydrin (99%) and 3-(2-methylphen-
oxy)-propane-1,2-diol (98%) were purchased from Alfa
Aesar1; 3-chloropropane-1,2-diol (991%), phenol (991%),
o-cresol (991%), 2-ethylphenol (99%) were purchased from
Acros Organics1; 2-propylphenol (97%), 2-isopropylphenol
(98%), 2-tert-butylphenol (99%), and 2-allylphenol (98%),
were purchased from Lancaster; methyl-tert-butyl ether
(MTBE, 99.5%) was obtained from Fisher Scientific.
Scheme 1. Synthesis of the chiral ortho-substituted phenyl glycerol
ethers 1a-g.
cally active substances, including registered drugs like
fungicide chlorphenesin,8 expectorant guaifenesin,8 mus-
cle relaxant mephenesin,8 belong to the series of aryl glyc-
erol ethers, possibly because of their similarity to lipids.
Thirdly, the terminal aryl glycerol ethers by simplest
chemical manipulations could be transformed to another
drugs, in particular to b-adrenoblockers of vast ArOꢀꢀCH2-
CH(OH)CH2ꢀꢀNHR family.9,10 Last, fourthly, it is rela-
tively easy to prepare both racemic and scalemic samples
of the aryl glycerol ethers needed for experiments.
Recently, we have studied the crystallization peculiar-
ities in the series of ortho-halogen and ortho-alkoxyphenyl
glycerol ethers.11,12 The present study is carried out to
examine the chirality-driven crystallization of a series of
ortho-substituted phenyl glycerol ethers, where the sub-
stituent is the hydrogen atom or a hydrocarbon moiety.
The availability of both enantiomers of 3-chloropropane-
1,2-diol 2 by Jacobsen kinetic hydrolytic resolution of a ra-
cemic epichlorohydrin13 makes it possible to obtain all the
series of glycerol ethers 1a–g through the use of a single
process (Scheme 1).
The purposes of this work are to study the melting of
the substituted phenyl glycerol ethers as enantiomers, as
racemates, and in some cases as mixtures of intermediate
compositions by means of DSC; to evaluate stabilities of ra-
cemic compounds under experiment; to compare the crys-
tallization peculiarities with the substitution pattern; and to
investigate the abilities of conglomerate forming com-
pounds to be resolved by entrainment procedure.
Preparation of 3-Aryloxypropane-1,2-diols
Racemic and enantiopure diols 1a–g were synthesized
by analogy with a published procedure14 from correspond-
ing phenol 3a–g and racemic or scalemic 3-chloropropane-
1,2-diols, rac-2 and scal-2, respectively. (2)-(R)-2 {[a]D20
26.4 (c 5, H2O)} and (1)-(S)-2 {[a]D20 16.1 (c 4.8, H2O)}
were prepared through Jacobsen kinetic hydrolytic resolu-
tion of rac-epichlorohydrin without modifications.13
General procedure. To the solution of corresponding
phenol 3a–g (10 mmol) in ethanol (6 ml) a solution of
NaOH (0.5 g, 12.5 mmol) in water (2 ml) was added and
the resulting mixture was stirred and heated under reflux
for 30 min. Then a solution of 3-chloropropane-1,2-diol 2
(1.3 g, 12 mmol) in ethanol (1 ml) was added, and the mix-
ture was further stirred and heated at reflux for 3 h. After
cooling, the mixture was concentrated under reduced
pressure followed by addition of water (6 ml) and extrac-
tion with CH2Cl2 (3 3 20 ml). The combined organic
layers were dried over anhydrous Na2SO4 and the solvent
was removed. The crude diols 1a–g were usually purified
by recrystallization from hexane; (S)-aryloxypropanediols
were obtained from (1)-(S)-2, and (R)-aryloxypropane-
diols were obtained from (2)-(R)-2.
MATERIALS AND METHODS
General
1
The NMR spectra (600 MHz for H and 150.864 MHz
for 13C) were recorded on a Bruker Avance-600 spectrome-
ter in CDCl3 with TMS or the signals of the solvent as the
internal standard. The IR spectra of the polycrystalline
samples of rac- and scal-compounds under investigations
rac-3-Phenoxypropane-1,2-diol, rac-1a. Yield: 1.3 g
in KBr pellets were recorded on a Bruker IFS-66v Fourier- (80%), mp 58–598C (Ref. 11: mp 56–608C). An analytical
transform spectrometer.
sample was obtained by slow crystallization of crude rac-1a
1
The X-ray single crystal experiments were carried out from a hexane-ether (1:1) mixture. H NMR, d: 2.55 (br.s,
on Bruker AXS Smart Apex II and Kappa Apex II diffrac- 2H, 2 OH), 3.75 (dd, J 5 11.7, 5.5 Hz, 1H, 1CH2O), 3.84
tometers. Powder X-ray diffraction (PXRD) data were col- (dd, J 5 11.7, 3.8 Hz, 1H, 1CH2O), 3.99–4.04 (m, 2H,
lected on a Bruker D8 Advance diffractometer equipped CH2O), 4.08–4.15 (m, 1H, CHO), 6.91 (dd, J 5 8.4, 1.0 Hz,
with an Vantec linear PSD, using graphite monochromated 2 H, C2A;r6 H), 6.98 (t, J 5 7.4 Hz, 1H, C4Ar H), 7.29 (td, J 5
3;5
Ar
7.4, 1.4 Hz, 2H, C H). 13C NMR, d: 63.83 (CH2OH), 69.28
˚
Cu Ka (k 1.54184 A) radiations, (40 kV, 40 mA).
Optical rotations were measured on a Perkin-Elmer (CH), 70.57 (CH2O), 114.71 (C3A;r5), 121.48 (CA4 r), 129.70
model 341 polarimeter. Throughout the paper the value of (C2A;r6), 158.55 (CA1 r).
Chirality DOI 10.1002/chir

1094
BREDIKHIN ET AL.
(R )-3-Phenoxypropane-1,2-diol, (R )-1a. Yield: 1.2 0.6 ml/min; tR (minor) 5 28.2 min, tR(major) 5 33.4 min].
g (71%), mp 68–698C (Ref. 15: mp 62.5–64.58C); [a]D20
–
1H NMR d: 2.63 (broad s, 1H, OH), 3.03 (broad s, 1H,
10.1 (c 1, EtOH), [a]D20 29.7 (c 1.0, MeOH) {Ref. 15: [a]D20 OH), 3.34–3.42 (m, 2H, ꢀꢀCH2ꢀꢀCH¼¼), 3.73 (dd, J 5 11.5,
29.5 (c 0.5, MeOH), 98.9% ee}, [a]D20 14.2 (c 1, hexane/ 5.5 Hz, 1H, CH2O), 3.81 (dd, J 5 11.4, 3.3 Hz, 1H, CH2O),
EtOH, 4:1), [a]D20 1 2.0 (c 1, MTBE); 99.6% ee [HPLC; 4.01–4.08 (m, 2H, CH2O), 4.11–4.19 (m, 1H, CH), 4.98–
column temperature 298C; eluent: water/isopropanol 5 5.05 (m, ¼¼CH2), 5.92–6.02 (m, 1H, ¼¼CH), 6.82 (d, J 5
3.75/1; flow rate 1.0 ml/min; tR(major) 5 7.4 min]; tR 8.1 Hz, 1H, C6Ar H), 6.92 (t, J 5 7.3 Hz, 1H, C4Ar H), 7.12–
(minor) 5 9.6 min]. NMR spectra were identical with 7.19 (m, 2H, C3;5 H). 13C NMR d: 34.58 (CH2ꢀꢀCH¼¼),
Ar
spectra of rac-1a.
63.87 (OCH2), 69.38 (OCH2), 70.78 (CH), 111.82 (C6Ar),
115.30 (CH¼¼CH2), 121.27 (C4Ar), 127.54 (CA5 r), 130.20
(C3Ar), 137.27 (CH¼¼CH2), 128.73 (C2Ar), 156.34 (CA1 r).
(S)-3-(2-Methylphenoxy)-propane-1,2-diol, (S)-
1b. Yield: 1.4 g (77%), mp 90–918C (Ref. 16: mp 91–
928C); [a]D20 219.3 (c 1.2, hexane/i-PrOH, 4:1) {Ref. 17:
rac-3-(2-Propylphenoxy)-propane-1,2-diol, rac-1e.
[a]D20 219.3 (c 0.9, hexane/i-PrOH, 4:1)}, [a]D20 212.8 Yield: 1.6 g (76%), mp 52.5–53.58C (hexane); (Ref. 19: mp
1
(c 1.1, MTBE); 99.8% ee [HPLC; column temperature 52–538C). H NMR, d: 0.97 (t, 3H, CH3, J 5 7.3 Hz), 1.61–
408C; eluent: water/isopropanol 5 3/1; flow rate 1.0 ml/ 1.65 (m, 2H, CH2CH3), 2.61 (t, J 5 7.3 Hz, 2H, CH2CH2),
1
min; tR(major) 5 8.3 min]. H NMR (600 MHz) d 5 2.20 2.80 (br.s, 2H, OH), 3.69 (dd, J 5 11.5, 5.8 Hz, 1H, CH2O),
(br.s, 1H, OH), 2.23 (s, 3H, CH3), 2.70 (br.s, 1H, OH), 3.78 3.78 (dd, J 5 11.5, 3.7 Hz, 1H, CH2O), 3.95–3.98 (m, 2H,
(dd, J 5 11.8, 6.1 Hz, 1H, 1CH2), 3.85 (dd, J 5 11.8, 3.7 CH2O), 4.03–4.07 (m, 1H, CH), 6.73 (d, J 5 7.8 Hz, 1H,
C6Ar H), 6.83 (dd, J 5 6.8, 7.3 Hz, 1H, CA4 r H), 7.04–7.07 (m,
Hz, 1H, 1CH2O), 4.04–4.07 (m, 2H, CH2O), 4.11–4.14 (m,
1H, CH), 6.82 (d, J 5 7.9 Hz, 1H, CA6 r H), 6.88 (t, J 5 7.4 Hz, 2H, C3A;r5 H). 13C NMR, d: 14.07 (CH3), 23.07 (CH2), 32.24
1H, C4Ar H), 7.13–7.15 (m, 2H, CA3;r5 H).
(OCH2), 63.90 (CH2O), 69.07 (CH2O), 70.71 (CH), 111.46
(C6Ar), 121.04 (CA4 r), 126.94 (CA5 r), 130.06 (CA3 r), 131.08
(C2Ar), 156.22 (CA1 r).
(R)-3-(2-Methylphenoxy)-propane-1,2-diol, (R)-1b.
Yield: 1.4 g (77%), mp 90–918C; [a]D20 119.3 (c 1.2, hex-
ane/i-PrOH, 4:1) {Ref. 17: [a]D20 119.8 (c 0.9, hexane/i-
PrOH, 4:1)}, [a]D20 112.8 (c 1.1, MTBE); 99.9% ee [HPLC;
column temperature 408C; eluent: water/isopropanol 5 3/
1; flow rate 1.0 ml/min; tR(major) 5 7.1 min].
(R)-3-(2-Propylphenoxy)-propane-1,2-diol, (R)-1e.
Yield: 1.5 g (71%), mp 67–698C (hexane); [a]D20 114.9 (c 1,
hexane/ EtOH, 4:1); [a]D20 5 111.9 (c 1.0, MTBE); 99.9%
ee [HPLC; column temperature 288C; eluent: water/iso-
propanol 5 3/1; flow rate 1.0 ml/min; tR(major) 5 28.9
min, tR(minor) 5 33.2 min]. NMR spectra were identical
with spectra of rac-1e.
rac-3-(2-Ethylphenoxy)-propane-1,2-diol, rac-1c.
1
Yield: 1.5 g (76%), mp 51–538C (Ref. 18: mp 56–578C). H
NMR, d: 1.23 (t, J 5 7.6 Hz, 3H, CH3), 2.67 (2H, J 5 7.6
Hz, CH2CH3), 3.30 (br.s, 1H, OH), 3.53 (br.s, 1H, OH),
3.78 (dd, J 5 11.5, 6.3 Hz, 1H, 1CH2O), 3.87 (dd, J 5 11.5,
3.7 Hz, 1H, 1CH2O), 4.01–4.05 (m, 2H, CH2O), 4.03–4.07
(m, 1H, CH), 6.76 (d, J 5 8.4 Hz, 1H, C6Ar H), 6.85 (t,
J 5 7.3 Hz, 1H, C4Ar H), 7.06–7.09 (m, 2H, CA3;r5 H). 13C
NMR, d: 14.20 (CH3), 23.22 (CH2CH3), 63.93 (CH2OH),
69.03 (CH), 70.71 (CH2O), 111.41 (C6Ar), 121.18 (CA4 r),
126.90 (C5Ar), 129.09 (CA3 r), 132.66 (C2Ar), 156.14 (CA1 r).
rac-3-(2-Isopropylphenoxy)-propane-1,2-diol, rac-1f.
1
Yield: 1.8 g (88%), mp 80–818C (hexane). H NMR, d: 1.16
(d, 6H, CH3, J 5 6.8 Hz), 2.20 (br. s, 2H, OH), 3.20–3.25
(m, 1H, CH), 3.71 (dd, J 5 11.5, 5.6 Hz, 1H, CH2O), 3.80
(dd, J 5 11.5, 3.7 Hz, 1H, CH2O), 3.98–4.01 (m, 2H,
CH2OH), 4.05–4.09 (m, 1H, CH), 6.77 (d, J 5 8.1 Hz, 1H,
C6Ar H), 6.89 (t, J 5 7.3 Hz, 1H, CA4 r H), 7.08 (t, J 5 7.3 Hz,
1H, C5Ar H), 7.15 (d, J 5 7.3 Hz, 1H, CA3 r H). 13C NMR, d:
22.70 (CH3), 26.87 (CH), 63.89 (OCH2), 69.29 (CH2O),
70.64 (CH), 111.54 (C6Ar), 121.37 (CA4 r), 126.20 (CA5 r),
126.68 (C3Ar), 137.00 (C2Ar), 155.47 (CA1 r).
(S)-3-(2-Ethylphenoxy)-propane-1,2-diol, (S)-1c.
Yield: 1.3 g (68%), mp 68–708C; [a]D20 214.5 (c 1, hexane/
EtOH, 4:1); [a]D20 212.3 (c 1.0, MTBE); 99.9% ee [HPLC;
column temperature 298C; eluent: water/isopropanol 5 3/
1; flow rate 1.0 ml/min; tR(major) 5 18.1 min]. NMR spec-
tra were identical with spectra of rac-1c.
(R)-3-(2-Isopropylphenoxy)-propane-1,2-diol, (R)-
1f. Yield: 1.6 g (76%), mp 71–728C (petroleum); [a]D20
113.7 (c 1, hexane/i-PrOH, 4:1); [a]D20 110.7 (c 1.0,
MTBE) {Ref. 14: [a]D20 114.1 (c 1, hexane/i-PrOH, 4:1)};
99.0% ee [HPLC; column temperature 21.58C; eluent:
water/isopropanol 5 3.75/1; flow rate 0.8 ml/min; tR(ma-
jor) 5 67.2 min, tR(minor) 5 86.1 min]. NMR spectra were
identical with spectra of rac-1f.
(R)-3-(2-Ethylphenoxy)-propane-1,2-diol, (R)-1c.
Yield: 1.4 g (72%), mp 68–708C; [a]D20 111.9 (c 1.0,
MTBE); 99.3% ee [HPLC; column temperature 298C; elu-
ent: water/isopropanol 5 3/1; flow rate 1.0 ml/min; tR
(major) 5 15.6 min].
rac-3-(2-Allylphenoxy)-propane-1,2-diol, rac-1d.
Yield: 1.6 g (75%), mp 41–438C (hexane) (Ref. 9: mp 43–448C).
rac-3-(2-tert-Butylphenoxy)-propane-1,2-diol, rac-
1g. Yield: 1.4 g, 63%, mp 68–718C (hexane/benzene).
1
(S)-3-(2-Allylphenoxy)-propane-1,2-diol, (S)-1d. (Ref. 20: mp 568C). H NMR, d: 1.41 (s, 9H, CH3), 2.35
Yield: 1.4 g (70%), mp 57–588 (hexane); [a]D20 22.7 (c 0.6, (br.s, 2H, OH), 3.82 (dd, J 5 11.5, 6.0 Hz, 1H, CH2O), 3.92
EtOH); [a]D20 27.2 (c 1.0, MTBE); {Ref. 9: mp 47–498C; (dd, J 5 11.5, 3.7 Hz, 1H, CH2O), 4.07–4.12 (m, 2H,
[a]D20 22.1 (c 2.8, EtOH)}; 99.9% ee [HPLC; column tem- OCH2), 4.21–4.24 (m, 1H, CH), 6.91 (d, J 5 7.8 Hz, 1H,
perature 298C; eluent: water/isopropanol 5 3/1; flow rate C6Ar H), 6.95 (dd, J 5 7.3, 7.8 Hz, 1H, C4Ar H), 7.20 (t, J 5
Chirality DOI 10.1002/chir

THREE TYPES OF CHIRALITY-DRIVEN CRYSTALLIZATION
1095
TABLE 1. Resolution by entrainment of rac-mephenesin 1b (500 cm3 H2O, 25 mg crystal seeds on every run,
crystallization temperature 258C 6 0.58C)
Operation amount of
(R)- and (S)-1ba, (g)
(R)- and (S)-1b obtained
Added amount
of (R,S)-1b, (g)
Resolution
time (min)
Run
(R)-1b
(S)-1b
Yield (g)
ee (%)b
YE (g)c
1
2
3
4
5
6
9.0d
1.80
1.64
1.48
1.53
1.60
5.50
4.63
5.41
4.68
5.39
4.63
4.50
5.37
4.59
5.32
4.61
5.37
95
80
75
80
75
85
(R) 1.82
(S) 1.66
(R) 1.50
(S) 1.55
(R) 1.62
(S) 1.54
97.0
95.4
98.5
92.7
95.0
97.0
1.77
1.60
1.47
1.48
1.56
1.50
aThe operation amounts in even runs 2–6 were calculated based on the results in odd runs 1–5, respectively.
bee: enantiomeric excess (HPLC).
cYE: Yield of enantiomer; YE (g) 5 [Yield (g) 3 ee (%)]/100 2 0.025 (seed weight).
dAdditional amount of (R)-1b 1.0 g.
7.3 Hz, 1H, C5Ar H), 7.32 (d, J 5 7.8 Hz, 1H, CA3 r H). 13C
NMR, d: 30.16 (CH3), 34.93 (CMe3), 64.12 (OCH2), 69.15
(CH2O), 70.95 (CH), 112.42 (C6Ar), 121.10 (CA4 r), 126.97
(C5Ar), 127.28 (CA3 r), 138.24(C2Ar), 157.34 (CA1 r).
Resolution of Racemic 3-(2-Ethylphenoxy)-propane-
1,2-diol (rac-1c) by Preferential Crystallization
Racemic diol rac-1c (1.0 g) and (S)-1c (0.1 g) was dis-
solved in 115 ml of water at 53–558C. The solution was
cooled to 258C and seeded with finely pulverized (S)-1c
(2.5 mg). After stirring the mixture for 70 min at 208C 6
0.58C, precipitated (S)-1c was collected by filtration (0.20
g after drying; 90% ee). The extra portion of rac-1c (0.2 g)
was then dissolved in the mother liquor at 558C; the result-
ing solution was cooled to 258C. After the addition of (R)-
1c (2.5 mg) as seed crystals to the solution, and stirring
the mixture for 80 min at 208C 6 0.58C, (R)-1c (0.18 g af-
ter drying; 89% ee) was collected by filtration.
(S)-3-(2-tert-Butylphenoxy)-propane-1,2-diol, (S)-
1g. Yield: 1.5 g (67%), [a]D20 210.5 (c 1, hexane/EtOH,
4:1); [a]D20 29.5 (c 1.0, MTBE). 99.3% ee [HPLC; column
temperature 218C; eluent: water/isopropanol 5 3/1; flow
rate 0.8 ml/min, tR(major) 5 50.5 min]. NMR spectra
were identical with spectra of rac-1g.
(R)-3-(2-tert-Butylphenoxy)-propane-1,2-diol, (R)-
1g. mp 94–958C (hexane); [a]D20 110.5 (c 1, hexane/
EtOH, 4:1); [a]D20 19.5 (c 1.0, MTBE). 99.7% ee [HPLC;
column temperature 218C; eluent: water/isopropanol 5 3/
1; flow rate 0.8 ml/min, tR (major) 5 46.3 min].
X-ray Diffraction Experiments for 3-(2-tert-
Butylphenoxy)-propane-1,2-diols, 1g
The X-ray diffraction data for crystals of compounds rac-
1g and (S)-1g were collected on a Bruker AXS Smart
Apex II and Kappa Apex II [(S)-1g] diffractometers in the
x and j-scan modes, using graphite monochromated Mo
Resolution of Racemic 3-(2-Methylphenoxy)-propane-
1,2-diol (Mephenesin, rac-1b) by Preferential
Crystallization (Entrainment)
˚
˚
Racemic mephenesin rac-1b (9.0 g) and (R)-1b (1.0 g) Ka k 0.71073 A (rac-1g) and Cu Ka k 1.54184 A [(S)-1g]
was dissolved in 500 ml of water at 53–558C. The solution radiations. Data were corrected for the absorption effect
was cooled to 288C and seeded with finely pulverized (R)- using SADABS program.21 The structures were solved by
1b (25 mg). After stirring the mixture for 95 min at 258C direct method and refined by the full matrix least-squares
6 0.58C, precipitated (R)-1b was collected by filtration using SHELXTL22 and WinGX23 programs. All nonhydro-
(1.82 g after drying; 97% ee) (Table 1, run 1). The extra gen atoms were refined anisotropically. The positions of
portion of rac-1b (1.80 g) was then dissolved in the mother hydrogen atoms were located from the Fourier electron
liquor at 558C; the resulting solution was cooled to 288C. density synthesis and were included in the refinement in
After the addition of (S)-1b (25 mg) as seed crystals to the the isotropic riding model approximation. Illustrations
solution, and stirring the mixture for 80 min at 258C 6 were made using PLATON.24
0.58C, (S)-1b (1.66 g after drying; 95% ee) was collected by
Crystallographic data (excluding structure factors) for
filtration (run 2). Further resolution was carried out at the structures reported in this article have been deposited
258C 6 0.58C by adding amended amounts of rac-1b to the in the Cambridge Crystallographic Data Centre as supple-
filtrate in a manner similar to that described above. The mentary publication nos. CCDC 691121 for rac-1g, and
detailed conditions are given in Table 1. A high degree of 691122 for (S)-1g. Copies of the data can be obtained free
enantiomeric purity of collected diols can be achieved by of charge upon application to the CCDC (12 Union Road,
simple recrystallization. For example, a portion of (R)-1b Cambridge CB2 1EZ U.K. Fax: (internat.) 144-1223/336-
(1.62 g, 95% ee) was dissolved in the boiling hexane (64 033; E-mail: deposit@ccdc.cam.ac.uk).
ml). After cooling the solution to 228C, the crystallized
(R)-1b was collected by filtration (yield 1.24 g; 99.5% ee).
Room-temperature powder X-ray diffraction data were
collected in the Bragg–Brentano mode with a flat-plate
Chirality DOI 10.1002/chir

1096
BREDIKHIN ET AL.
sample. The sample was lightly ground and loaded into
a standard sample holder, which was kept spinning
(30 rpm) throughout the data collection. Patterns were
recorded in the 2y range between 38 and 508, in 0.0088
steps, with a step time of 2 sec. Five powder patterns were
collected and summed for each sample.
Crystallographic Data for ( S)-3-(2-tert-Butylphenoxy)-
propane-1,2-diol, ( S)-1g
The single crystal picked out from the crop of the afore-
mentioned scal-1g with a specific rotation value of [a]D20
29.5 (c 1.0, MTBE) was used for single crystal experi-
ment. Formula C13H20O3, colorless prism of dimensions
0.35 3 0.24 3 0.14 mm3, formula weight 224.29 g mol21
,
monoclinic, a 5 26.5455(8), b5 6.9990(3), c5 16.6275(6)
3
˚
˚
A, b 5 124.359(2)8, V 5 2550.23(16) A , T 5 293 K, space
group C 2 (No. 5), Z 8, l(Cu-Ka) 6.57 cm21, F(000) 5 976,
dcalc 5 1.168 g cm23, 4917 reflections measured, 2106
unique (Rint 5 0.0673), 1961 observed with Fo > 4r (Fo),
378 parameters. The final R1(F2) was 0.0480 (>2rI) and
wR2(all data) 5 0.1291. Goodness-of-fit on F2 was 1.025,
3
˚
largest diff. peak and hole are 0.203 and 20.194 eꢂA . The
final value 20.10(3) of Flack enantiopole parameter25,26
proved the expected S absolute configuration.
The finely pulverized portion from the same bulk sam-
ple was used for PXRD experiments.
Crystallographic Data for rac-3-(2-tert-Butylphenoxy)-
propane-1,2-diol, rac-1g
The single crystal picked out from the crop of freshly
crystallized rac-1g was used for determination of crystal
and molecular structure of primary racemate. The eutectic
character of the crop and the metastability of this crystal
modification (see text) are the important reasons preclud-
ing the high quality crystal growth. Nevertheless the
experiment with this crystal enables us to evaluate some
important features of primary racemate structure. Formula
C13H20O3, colorless prism of dimensions 0.70 3 0.46 3
0.12 mm3, formula weight 224.29 g mol21, triclinic, a 5
˚
6.6876(11), b 5 14.659(2), c 5 14.648(2) A, a 5 70.083(2)8,
3
˚
b 5 76.777(2)8, g 5 76.785(2)8, V 5 1296.3(4) A , T 5
293 K, space group P 1 (No. 1), Z 4, l(Mo Ka) 0.80 cm21
,
dcalc 5 1.149 g cm23, F(000) 5 488, 10,227 reflections
measured, 9387 unique (Rint 50.0173), 3565 observed
with Fo > 4r (Fo), 556 parameters. The final R1(F2) was
0.0900 (>2rI) and wR2 (all data) 5 0.2970. Goodness-of-fit
on F2 was 0.927, largest diff. peak and hole are 0.343 and
3
˚
20.223 eꢂA .
The finely pulverized portion from the same bulk sam-
ple was used for PXRD experiments.
RESULTS AND DISCUSSION
IR Spectroscopy Investigations of the Racemic and
Scalemic Samples of the Compounds 1a–g
Fig. 1. IR spectra of the crystalline samples 1a–g. Red curves—race-
mates, blue curves—scalemates, black curves are differential curves (see
text). [Color figure can be viewed in the online issue, which is available at
www.interscience.wiley.com.]
For initial evaluation of the crystallization type of our
compounds, we compared the IR spectra of the racemic
Chirality DOI 10.1002/chir

THREE TYPES OF CHIRALITY-DRIVEN CRYSTALLIZATION
1097
TABLE 2. DSC measured melting point (T f ) and enthalpy of fusion (DH f ) of racemic (low index R) and enantiopure
(low index A) compounds 1a–f and calculated thermodynamic characteristics for these substances, calculated and
measured eutectic (low index eu) fusion temperature and eutectic enantiomeric composition (xÐmole fraction
of a predominant enantiomer)
TAf
(8C)
TRf
(8C)
DHfA
DHfR
Tefu
calc. (8C)
,
Tefu
exp. (8C)
,
DSm1
DG0
Comp.
(KJ mol21
)
(KJ mol21
)
xeu
(J K21 mol21
)
(J mol21
)
1a
1b
1c
1d
1e
1f
68.3
91.0
68.9
58.0
67.3
71.8
58.5
70.6
50.9
41.7
53.3
80.5
31.8
34.4
35.0
28.8
31.9
30.6
28.0
32.2
34.8
27.8
29.5
31.5
56.5a
70.1b
50.7b
41.3a
52.6a
67.5a
56.3
70.6
50.9
41.7
53.0
67.7
0.68
0.50
0.50
0.58
0.60
0.87
2.59
5.42
5.68
4.44
3.89
2994
251
221
2389
2561
22765
22.22
aCalculated as intersection for Schro¨der-Van Laar and Prigogine-Defay curves branches.
bCalculated as intersection for two Schro¨der-Van Laar curves branches.
and highly enantiomerically enriched crystalline samples line type for other investigated compounds needs further
of 1a–g in KBr pellets, since the IR spectra of an optically consideration.
active and the racemic form should be identical for the
normal conglomerate formative compounds. To substanti-
Thermochemical Investigations and Phase
ate this comparison, the spectra were subjected to a proce-
Behavior of the Compounds 1a–f
dure of normalization and baseline correction. For this
The results obtained for the temperature (Tf) and the
purpose, coefficients that minimize the difference ln(As) –
enthalpy of fusion (DHf ) of the pure enantiomers (low
[a0 1 a1m 1 ln(Ar)a2], where ln(As) and ln(Ar) are the
index A) and the pure racemates (low index R) of aryl
glycerol ethers 1a–f are presented in Table 2.
extinctions (transmission logarithms) of the scalemic and
racemic samples, respectively; m is the IR radiation fre-
quency corresponding to A, and an are the desired regres-
sion coefficients, were selected by the least-squares
method. It was reasonable to introduce the regression
terms a1m to correct the spectral differences caused by the
nonspecific (not related to particular absorption bands)
interaction of IR radiation with matter (probably, by radia-
tion scatter on heterogeneities of the sample). It should be
noted that the use of polynomials of higher powers (quad-
ratic and cubic) for the generation of differential spectra
does not improve the statistical parameters characterizing
regression. The ratio between the mean-square deviation
In the idealized case, to construct the liquidus of the bi-
nary melting phase diagram in the region of a pure compo-
nent fusion, it is enough to have data on the fusion temper-
ature and enthalpy of fusion for an enantiomerically pure
sample. In this case the liquidus line is described by
Schro¨der-Van Laar equation, which is usually used in the
simplified form3
!
DHAf
1
TAf
1
Tf
ln x ¼
ꢀ
ð1Þ
R
of the differential curves and the averaged mean-square where x is the mole fraction of one of the enantiomers in
deviation of spectral curves for the racemate and scale- the mixture (the mole fraction of another enantiomer is x0
mate, that is, the ratio of error to variation (%), was used 5 1 – x); R is the universal gas constant (R 5 8.3170 J K21
as a quantitative characteristic for differential curves, and mol21). The liquidus lines of the phase diagrams for the
this namely quantities are cited in Figure 1.
compounds 1a–f, which were calculated by eq. 1 using ex-
Figure 1 shows good coincidence between the pairs of perimental data for enantiomerically pure diols, are pre-
spectra for compounds 1b and 1c under visual compari- sented in Figure 2 (blue lines). The same figures show
son; the same is almost true for the allyl and normal propyl the experimentally obtained liquidus points drawn by
derivatives 1d,e, whereas the spectra of racemic and enan- circles. For samples of diols with an intermediate enantio-
tiopure crystalline samples for unsubstituted compound meric composition, the experimental DSC thermograms
1a, isopropyl, and tert-butyl substituted phenyl glycerol contain additional peaks with the almost constant tempera-
ethers 1f,g, differ noticeably. A similar pattern can be ture Tefu which corresponds to the melting point of the
observed for the differential curves: for compounds 1b eutectics. These experimental characteristics are put on
and 1c, and to a certain extent for compounds 1d,e, differ- the plot as solid circles too. It is significant that for the
ences between the spectra of the racemate and enantioen- compounds 1b and 1c experimental liquidus points,
riched sample are about the same level as instrumental including points for racemates, practically fall on the
background. At the same time they are rather substantial Schro¨der-Van Laar curves.
for compounds 1a,f,g. Thus, IR-test is consistent with the
The experimental data for racemates, TRf and DHRf , allow
conglomerate nature of mephenesin 1b; and the great to construct another type of a theoretical curve describing
probability exists that normal racemic conglomerate is liquidus line for congruent melting of binary molecular
also formed by ethyl derivative 1c. The question of crystal- compounds.
Chirality DOI 10.1002/chir

1098
BREDIKHIN ET AL.
In this case, the liquidus line obeys the Prigogine–Defay
equation, which is usually used in the simplified form3
!
2DHRf
1
TRf
1
Tf
ln 4xð1 ꢀ xÞ ¼
ꢀ
ð2Þ
R
The liquidus lines of the phase diagrams for the com-
pounds 1a–f, which were calculated by eq. 2 using experi-
mental data for racemic diols, are presented in Figure 2 by
red lines.
From Figure 2 it will be obvious that the end of fusion
temperatures (colored circles position) for compounds 1a,
d–f are accurately and completely described by combination
of theoretical curves of both types. Hence, complete melting
phase diagrams could be reproduced by solid liquidus lines
fragments with the addition of probable solidus lines (black
straight lines which are drawn parallel to the x-axis).
Inspecting the forms of so-built phase diagrams one can
suppose that the methyl and ethyl substituted phenyl glyc-
erol ethers 1b,c crystallize as racemic conglomerates,
whereas the unsubstituted phenyl glycerol ether 1a and
its allyl and propyl derivatives 1d–f forms solid racemic
compounds. There are other points in favor of this distri-
bution of the crystallization types, we mean the calculated
values of entropy of mixing for the liquid enantiomeric
compounds, DSm1 , and the free energy of formation of race-
mic compounds in the solid state, DG0.
Based on a thermodynamic cycle involving the solid and
liquid phases of the enantiomers and racemic species, for-
mulas for the DG0 and DS1m calculations were proposed by
Grant and coworkers.5
ðTRf ꢀ TAf ÞDHAf
DG0Tf ¼ ꢀ
ꢀ TRf R ln2 ðif TRf ꢀ TAf < 0Þ ð3aÞ
TAf
R
ðTRf ꢀ TAf ÞDHRf
DG0Tf ¼ ꢀ
ꢀ TAf R ln2 ðif TRf ꢀ TAf > 0Þ ð3bÞ
TRf
A
DHAf DHRf ꢀ DHAf
DHRf
TRf
TRf
TAf
DSlm
¼
ꢀ
ꢀ
ln
ð4Þ
TAf
TRf ꢀ TAf
The values of DSm1 and DG0 calculated by the formulas
(3) and (4) are presented in Table 2. The entropy of mix-
ing for enantiomers 1b and 1c in the liquid state is equal
to 5.42 and 5.67 J K21 mol21, which is slightly less but
close to the value of 5.75 J K21 mol21 (Rln2) for an ideal
conglomerate. The near zero value for DG0 also points on
the same feature of chiral 1b and 1c.3,5 The relatively
high negative value for DG0 for unsubstituted and isopro-
pyl derivatives 1a and 1f is a good diagnostic for a stable
racemic compounds formation in the crystalline state.5
The intermediate DSm1 and especially DG0 values for diols
1d and 1e exclude a conglomerate formation, and are
compatible with the assumption of (rather unstable) race-
mic compound formation.
Fig. 2. Experimental (circles) points and calculated (solid lines frag-
ments) binary melting phase diagrams for compounds 1a–f. [Color figure
can be viewed in the online issue, which is available at www.interscience.
wiley.com.]
Further information about the compounds under investi-
gation could be picked up by inspection of the binary phase
diagram characteristics. The case at hand is the composition
of the eutectic. The last is the special point on the binary (or
Chirality DOI 10.1002/chir

THREE TYPES OF CHIRALITY-DRIVEN CRYSTALLIZATION
1099
ternary) phase diagram where the three phases, liquid (melt (one cycle) are sufficient to obtain from 1.2 g racemate
or solution) and two solid (either two individual enantiomers (R)- and (S)-1c samples of about 0.2 g each. The quality of
or enantiomer and racemic compound) exist in equilibrium. both specimens are good enough, and making sure that
If the enantiomeric excess of the eutectic point (eeeu) is the compound 1c is quite capable of preferential crystalli-
exactly equal to zero, we are dealing with a conglomerate zation we made no attempts to improve sample enantio-
forming compound. For this compound the direct resolution purity and to optimize experimental conditions for the
approaches could be realized and enantiomeric composition resolution.
of any scalemic (nonracemic) sample could be enriched to
For monitoring the entrainment abilities of mephenesin
any desirable degree by fraction crystallization. If the eutec- 1b during seed-induced crystallization of oversaturated
tic ee lies between 0 and 1 (0 < eeeu < 100%), we are deal- slightly nonracemic water solutions we have used mother
ing most likely with a racemic compound forming sub- liquor enantiomeric composition. Unfortunately, the opti-
stance. For this system, the standard direct resolution is cal rotation of water mephenesin solutions is too low to
impossible, but if the ee of the starting sample is lower than use polarimetry for these purposes. So we made use of chi-
eeeu, the predominant enantiomer can be enriched in the liq- ral liquid chromatography for monitoring this process.
uid phase (mother liquor) to the eutectic ee extent, whereas During the course of crystallization we have collected the
the precipitate will tend to be racemate. If the solution of the aliquots (10 ll) of mother liquor at regular intervals. The
starting material has an ee higher than the eeeu, a pure pre- aliquots were mixed with 1 ml of PriOH, and portions of
dominant enantiomer or mixtures of enantiomers with an ee 20 ll of the solutions were analyzed with the Chiralcel AD-
higher than the eutectic ee can be crystallized out. Equilib- RH column. All investigated solutions of mephenesin of
rium crystallization of the sample with ee 5 eeeu could not different initial concentrations but of equal starting enantio-
change its enantiomeric composition in either the precipitate meric enrichment behave uniformly in the main features.
or in the filtrate form.
During the crystallization process primary enantiomeric
The eutectic compositions for the glycerol ethers 1a–f excess of solute (ꢁ9%) decreased to zero, changed sign, and
are presented in Table 1. From this it follows that the sam- mounted to initial absolute value (even exceeded this value
ple could be enantioenriched by crystallization provided for high initial concentrations). The mother liquor enantio-
that an enantiomeric excess of it is about or more than meric composition (and the composition of deposited crys-
74% (1f), 36% (1a), and 20% (1e). For the allyl derivative tals as well) then asymptotically tended to racemic. Varying
1d the ee ꢃ 6% is sufficient for enantioenrichment, and at initial concentration, temperature gap, and crystallization
last for methyl and ethyl derivatives a sample could be ra- time we have adopted the conditions for the satisfactory re-
cemic for the purpose; in other words, the compounds 1b solution of racemic mephenesin. An example of successful
and 1c could be potentially resolvable by direct methods. resolution of near racemic 1b is illustrated in Table 1.
We have particularly inspected this possibility.
Over the course of the resolution a supersaturated solu-
tion of rac-1b, including a small excess of (R)-1b, was pre-
pared by heating, and then cooling to 258C. A small
amount of seed crystals of (R)-1b was added and the
stirred solution was allowed to crystallize for about 85 min.
The weight of (R)-1b obtained after filtration was more
than the common weight of the initial excess of the (R)-
Resolution of Racemic 3-(2-Methylphenoxy)-propane-
1,2-diol (Mephenesin, rac-1b) and 3-(2-ethylphenoxy)-
propane-1,2-diol (rac-1c) by Entrainment Procedure
The entrainment effect, that is the preferential crystalliza-
tion of an enantioenriched crop induced by seeding with
enantiopure crystals of the oversaturated solution of a con- enantiomer and seed added. rac-1b was added to the
mother liquor in order that the overall quantity of 1b in
glomerate forming (almost) racemic chiral compound, was
first discovered by Pasteur’s student Gernez.27,28 Racemates
the solution could be recovered. The mixture was heated
resolution by entrainment4 is a gratifying labor since success until the solid was completely dissolved and then cooled
to 258C. After the solution had been seeded with (S)-1b
and stirred for about 85 min, precipitated (S)-1b was col-
allows one to obtain easily both enantiomers without resort-
ing to any enantiopure auxiliaries. But even in the case of
conglomerate forming substances it is not always easy to lected in a similar manner. The cycle was repeated several
times.
As evident from Table 1, three cycles (six runs) of
entrainment resolution using ordinary laboratory equip-
ment and no resolving agents are sufficient enough to
obtain from ꢁ17 g of rac-1b (R)- and (S)-1b samples of
about 5 g each. The quality of each nonracemic specimen
use the benefits of a spontaneous resolution. Coquerel and
coworkers4,29 are of the opinion that almost half of conglom-
erate forming compounds would demonstrate poor entrain-
ment characteristics.
Discovering two new conglomerates in the family
of aryl glycerol ethers we decided to examine their
abilities for preferential resolution. This task was all the is good enough, and single recrystallization is sufficient to
obtain (R)- and (S)-1b diols with enantiomeric excess
higher than 99%.
more interesting since rac-3-(2-methylphenoxy)-propane-1,
2-diol, rac-1b is a well-known skeletal muscle relaxant, me-
phenesin.8 We used water for the resolution as the most
cheap and ‘‘green’’ solvent, and by analogy with our own
successful effective entrainment procedure for another chi-
ral drug, guaifenesin.10
Crystallization Features of 3-(2-tert-Butylphenoxy)-
propane-1,2-diol 1g
Both racemic (Fig. 3, curve 1, Fig. 4, curve 1) and enan-
In the case of ethyl derivative 1c we have just demon- tiopure (Fig. 4, curve 5) polycrystalline samples of 3-(2-
strated entrainment effect. As shown earlier, two runs tert-butylphenoxy)-propane-1,2-diol 1g obtained by crystal-
Chirality DOI 10.1002/chir

1100
BREDIKHIN ET AL.
crystals, and is formed by some other metastable (with
respect to a homochiral) solid phase(s).
It can be inferred that all the facts under discussion are
understood by reference to so-called anomalous race-
mate3,31 formation during the (solution) crystallization of
the racemic or the intermediate composition samples of
1g. According to Jacques et al.,3 anomalous racemate is an
addition molecular compound in which crystals contain
two enantiomers with a stoichiometry differing from the
usual 1:1 ratio. We have anticipated that the above dis-
cussed ‘‘primary racemate’’ of tert-butylphenyl glycerol
ether 1g is in fact the racemic conglomerate formed by
enantiomeric crystals of such anomalous racemate. To all
appearance, despite the fact of less thermodynamic stability,
the formation of this crystal modification is kinetically favor-
able. Therefore, during the racemate crystallization from a
solution only the ‘‘anomalous conglomerate’’ (this is a more
accurate expression for this case and the other examples of
the sort than ‘‘anomalous racemate’’) precipitates.
Fig. 3. DSC traces for the racemic samples of 3-(2-tert-butylphenoxy)-
propane-1,2-diol 1g; 1—‘‘primary racemate’’ obtained by solution crystalli-
zation of racemic 1g; 2—the same sample after several weeks aging; 3—
artificial recemate obtained by mixing of approximately equal amounts of
(R)- and (S)-1g.
If a solution contains an excess of one of enantiomers,
the precipitate, alongside with anomalous conglomerate,
lization from the solutions of corresponding composition contains individual enantiomer crystals. The presence of
this phase manifests itself by appearance of the third high
temperature peak on the DSC traces (Fig. 4). However,
the possibility of the stable enantiopure solid phase gener-
are melted reproducibly giving the single narrow DSC
traces with good linear fronts. This fact provides a way of
estimating the temperature and the enthalpy of fusion for
the pure enantiomer and the pure racemate of tert-butyl ation from the metastable phase(s) immediately in the
derivative 1g as Tf 5 94.58, DHf 5 32.2 KJ mol21, Tf
course of thermoscanning must not be ruled out. For the
high enantiomeric purity samples (x > 0.8) the anomalous
perimental characteristics allows reconstructing Schro¨der- racemate melting peak virtually goes to level (Fig. 4, curve
A
A
R
5 67.98C, DHf 5 19.3 KJ mol21. The first pair of the ex-
R
4). Evidently in this case two crystal phases present in the
sample, namely, one enantiomorph of the anomalous race-
5 72.28C. This mate and a pure enantiomer.
Van Laar branches of the phase diagram and estimating
(as the position of their intersection) the fusion tempera-
ture of hypothetical conglomerate, Tf
congl.
Bearing all these facts in mind we embarked on a series
value is sufficiently above the experimental one for race-
of control experiments using X-ray diffraction approaches.
mate. In the control experiment the racemic sample pre-
pared by mixing of (almost) equal amounts of the pure First of all we have established the molecular and crystal
structure of scal-1g crystallized from an enantiopure (by
HPLC) sample of the tert-butylphenyl glycerol ether by
enantiomer crystals has begun melting at near theoretical
temperature Tf
(Fig. 3, curve 3). After several weeks
congl
aging of the sample of ‘‘primary racemate’’ (obtained by
direct crystallization from a racemate solution) at an ambi-
ent temperature, its DSC melting trace splits, and a higher
melting component could be observed (Fig. 3, curve 2),
while no changes were detected in the melting of pure
enantiomer samples of the same age.
The anomalous behavior of the ‘‘primary racemate’’ sets
one thinking about additional metastable crystal forms in
the system under investigation.30 The shapes of the DSC
traces for intermediate enantiomeric purity samples pre-
pared by drying the solutions of proper composition (Fig.
4) provide support for this conclusion.
Thermograms for the mixtures of intermediate composi-
tions (Fig. 4, curves 2–3) show simultaneous presence of
three melting peaks. The low temperature one is identical
by appearance and position to the melting of ‘‘primary
racemate.’’ The peak is traceable in a wide range of com-
positions; the peak characteristic temperature is therewith
kept constant. This suggests that no individual compound
but some eutectic mixture melting is the thermoinitiated
process behind the low temperature peak appearance.
This eutectic is not coincident with the above studied tra-
Fig. 4. DSC traces for the varying composition samples of 3-(2-tert-
butylphenoxy)-propane-1,2-diol 1g; x 5 0.6 (1); x 5 0.7 (2); x 5 0.8 (3);
x 5 0.9 (4); x 5 1.0, pure enantiomer (5). Here x—mole fraction of a pre-
ditional racemic conglomerate formed by enantiopure dominant enantiomer.
Chirality DOI 10.1002/chir

THREE TYPES OF CHIRALITY-DRIVEN CRYSTALLIZATION
1101
opposite configuration than the fourth one has. There are
no crystallographic reasons to decide between these two
possibilities. We have yielded to chiral P1 group based on
the foregoing thermodynamic arguments. The case of
achiral group presumes the existence of an individual race-
mic compound in the solid primary racemate and conse-
quently the lowering of a fusion temperature of the sample
after the enantiopure component was added owing to new
eutectic formation. This is not the case as it was estab-
lished in the experiments (see Fig. 4). On the contrary the
case of noncentrosymmetrical P1 group presumes the
crystallization of the rac-1g primary racemate as a mixture
of enantiomorphous crystals having triple excess of one or
another enantiomer. This mixture constitutes a eutectic by
itself, and owing to this fact has a possibility to keep a con-
stant melting temperature with the advent of additional
component.
Fig. 5. ORTEP drawing of the independent part of the unit cell for (S)-
1g crystal. Displacement ellipsoids are drawn at the 50% probability level;
H atoms are represented by circles of arbitrary size.
Figure 6 demonstrates an independent part of the P1
unit cell for the crystal picked up from the primary race-
mate bulk sample. The calculated PXRD pattern for this
cell as well as for an experimental one for the freshly crys-
single crystal X-ray analysis. Leaving out the details, we
will point to Figure 5 where the independent part (two
molecules) of the unit cell of scal-1g is depicted. The pow-
der X-ray diffractogram has been simultaneously obtained
from the bulk sample of the scal-1g (Fig. 7d); the experi-
mental PXRD pattern has coincided completely with the
theoretical one (Fig. 7c).
When a single crystal picked up from the bulk sample
of ‘‘primary racemate’’ was examined by X-ray diffraction
analysis two possibilities for interpretation of the experi-
mental diffraction pattern were aroused. The automatically
offered one fitted with achiral space group Iba2 (No. 45)
with a sufficiently disordered position of all OH groups
and low refinements parameters. The disorder could be
diminished by lowering the crystal symmetry as far as P1
space group. In this case the unit cell consists of four inde-
pendent molecules with the first three of them having an
Fig. 7. Powder X-ray diffractograms for compound 1g: (a) calculated
from single crystal X-ray data based upon P1 unit cell parameters of pri-
mary racemate; (b) experimental pattern of freshly crystallized primary
racemate; (c) calculated from single crystal X-ray data for enantiopure
crystal; (d) experimental pattern for bulk enantiopure sample; (e) experi-
mental pattern of bulk racemic sample after 2 months of aging.
Fig. 6. ORTEP drawing of the independent part of the unit cell for rac-
1g crystal. Displacement ellipsoids are drawn at the 30% probability level;
H atoms are represented by circles of arbitrary size.
Chirality DOI 10.1002/chir

1102
BREDIKHIN ET AL.
2. Perez-Garcia L, Amabilino DB. Spontaneous resolution, whence and
whither: from enantiomorphic solids to chiral liquid crystals, mono-
layers and macro- and supramolecular polymers and assemblies.
Chem Soc Rev 2007;36:941–967.
tallized bulk sample of rac-1g are reproduced in Figure 7
(a and b correspondingly); both patterns are coincident
with one another. Figure 7e demonstrates the PXRD pat-
tern for the same rac-1g bulk sample after ꢁ2 months of
aging under ambient temperature. It will readily be seen
that this curve coincides almost completely with the above
discussed curve (Fig. 7d) for the enantiopure sample. And
this is direct evidence of solid phase change from metasta-
ble primary crystals (anomalous conglomerate) to stable
normal conglomerate formed by enantiopure components.
3. Jacques J, Collet A, Wilen SH. Enantiomers, racemates and resolu-
tions. Malabar, FL: Krieger Publishing; 1994. 447 p.
4. Coquerel G. Preferential crystallization. Top Curr Chem 2007;269:1–
51.
5. Li ZJ, Zell MT, Munson EJ, Grant DJW. Characterization of race-
mic species of chiral drugs using thermal analysis, thermody-
namic calculation, and structural studies. J Pharm Sci 1999;88:
337–346.
6. Ho¨hne GWH, Hemminger WF, Flammersheim H-J. Differential scan-
ning calorimetry. Berlin-Heidelberg-New York: Springer Verlag; 2003.
298 p.
CONCLUSIONS
7. Avalos M, Babiano R, Cintas P, Jimenez JL, Palacios JC. Symmetry
breaking by spontaneous crystallization—is it the most plausible
source of terrestrial handedness we have long been looking for?—A
reappraisal. Orig Life Evol Biosphere 2004;34:391–405.
The seven aryl glycerol ethers 1a–g studied in this
work have very close chemical structure and differ only in
the uniformly positioned (ortho phenoxyl position) far
removed from the chiral centers substituents. All the sub-
stituents, hydrogen atom, alkyl, or allyl groups, are alike
in polarity, are not capable of strong hydrogen bond forma-
tion, and are distinguishable mainly in size and conforma-
tional properties. Nevertheless these seemingly second-
order distinctions were found to be sufficient for consider-
able changes in the crystallization characteristics of the
different racemic members of the family.
Thus, three racemates out of the series of seven, H-, n-
Pr-, and i-Pr-derivatives 1a,e,f, crystallize as typical race-
mic compounds if not identical in thermodynamic charac-
teristics. Allyl derivative 1d forms rather unstable racemic
compound. It is not improbable that the change of crystal-
lization conditions will change the crystallization type in
this case. Two substances 1b and 1c turn out to be typical
conglomerates. The property of their spontaneous resolu-
tion could be used as the basis for an effective production
of these compounds in a single enantiomer form. This is
particularly of importance because one of these sub-
stances, 1b, is the registered drug mephenesin.
Finally, the series is closed by the tert-butyl derivative
1g, a racemate of which manifests rare if not unique types
of spontaneous resolution. It seems reasonable to say that
the metastable form, resulting during solution crystalliza-
tion, and the stable form, to which metastable one trans-
forms as times goes by, both are the eutectic mixtures of
scalemic crystals. Metastable eutectic, i.e. the conglomer-
ate of enantiomorphous crystals enriched with one or
another enantiomer, arising as a result of spontaneous re-
solution of the anomalous racemate, turns into the eutectic
of enantiopure crystals.
8. O’Neil MJ, editor. The Merck Index, 14th ed. Whitehouse Station, NJ:
Merck and Co; 2006.
9. Bredikhina ZA, Savel’ev DV, Bredikhin AA. Cyclic sulfites—the key
intermediates in the synthesis of 1-alkylamino-3-aryloxypropan-2-ols
from glycidol. Russ J Org Chem 2002;38:213–219.
10. Bredikhina ZA, Novikova VG, Zakharychev DV, Bredikhin AA. Solid
state properties and effective resolution procedure for guaifenesin, 3-
(2-metoxyphenoxy)-1,2-propanediol. Tetrahedron: Asymmetry 2006;
17:3015–3020.
11. Zakharychev DV, Lazarev SN, Bredikhina ZA, Bredikhin AA.
Crystallization of chiral compounds 3. 3-Phenoxypropane-1,2-diol and
3-(2-halophenoxy)propane-1,2-diols. Rus Chem Bull Int Ed 2006;55:
230–237.
12. Bredikhin AA, Bredikhina ZA, Zakharychev DV, Konoshenko LV.
Spontaneous resolution among chiral glycerol derivatives: crystalliza-
tion features of ortho-alkoxysubstituted phenyl glycerol ethers. Tetra-
hedron: Asymmetry 2007;18:1964–1970.
13. Schaus SE, Brandes BD, Larrow JF, Tokunaga M, Hansen KB, Gould
AE, Furrow ME, Jacobsen EN. Highly selective hydrolytic kinetic re-
solution of terminal epoxides catalyzed by chiral (salen)CoIII com-
plexes. Practical synthesis of enantioenriched terminal epoxides and
1,2-diols. J Am Chem Soc 2002;124:1307–1315.
14. Egri G, Kolbert A, Balint J, Fogassy E, Novak L, Poppe L. Baker’s
yeast mediated stereoselective biotransformation of 1-acetoxy-3-arylox-
ypropan-2-ones. Tetrahedron: Asymmetry 1998;9:271–283.
15. Kitaory K, Furukawa Y, Yoshimoto H, Otera J. CsF in organic
synthesis. Regioselective nucleophilic reaction of phenols with oxir-
anes leading to enantiopure b-blockers. Tetrahedron 1999;55:14381–
14390.
16. Bredikhin AA, Bredikhina ZA, Lazarev SN, Savel’ev DV. Systematic
search for conglomerates among glycerol aromatic monoethers: guai-
fenesin and mephenesin are the cases. Mendeleev Commun 2003;
13:104–105.
17. Theil F, Weidner J, Ballschuh S, Kunath A, Schick H. Kinetic re-
solution of acyclic 1,2-diols using a sequential lipase-catalyzed
transesterifications in organic solvents. J Org Chem 1994;59:388–
393.
Thus, in this work the queer character of a spontaneous
resolution has been demonstrated once again, the number
of known conglomerate forming substances has been
enlarged, and the uncommon variant of spontaneous reso-
lution of metastable anomalous racemate has been found
out.
18. Lambooy JP. Physiologically active substituted a-glyceryl phenyl
ethers. J Am Chem Soc 1951;73:349–350.
19. Yale HL, Pribyl EJ, Braker W, Bergeim FH, Lott WA. Muscle-relaxing
compounds similar to 3-(o-toloxy)-1,2-propanediol. I. Aromatic ethers
of polyhydroxy alcohols and related compounds. J Am Chem Soc
1950;72:3710–3716.
20. Auzou G, Rips R, Derappe C, Peyroux J. Anxiolytic activity derived
from a myorelaxant structure. tert-Butylphenol derivatives. Eur J Med
Chem Chim Ther 1974;9:548–554.
LITERATURE CITED
21. Sheldrick GM. SADABS, Program for empirical X-ray absorption cor-
rection. Bruker-Nonius, 2004.
1. Pasteur L. Sur les relations qui peuvent exister entre la forme cristal-
line, la composition chimique et le sens de la polarisation rotatoire.
Ann Chim Phys 1848;24:442–460.
22. Sheldrick GM. SHELXTL v.6.12, Structure determination software
suite, Bruker AXS, Madison, WI, 2000.
Chirality DOI 10.1002/chir

THREE TYPES OF CHIRALITY-DRIVEN CRYSTALLIZATION
1103
23. Farrugia LJ. WinGX 1.64.05. An integrated system of Windows pro-
grams for the solution, refinement and analysis of single crystal X-ray
diffraction data. J Appl Crystalogr 1999;32:837–838.
28. Gernez. Verfahren zur Scheidung rechtsdrehend- und linksdrehend-
weinsaurer Salze. Ann Chem Pharm 1867;143:376.
29. Dufour F, Perez G, Coquerel G. A Priori assessment of the maximum
possible entrainment effect attainable during preferential crystalliza-
tion. The case of the simultaneous resolution of (6)-ephedrine and
(6)-mandelic acid. Bull Chem Soc Jpn 2004;77:79–86.
24. Spek AL. Single-crystal structure validation with the program PLA-
TON. J Appl Cryst 2003;36:7–13.
25. Flack HD. On enantiomorph polarity estimation. Acta Crystallogr Sect
A 1983;39:876–881.
30. Coquerel G. Review on the heterogeneous equilibria between con-
densed phases in binary systems of enantiomers. Enantiomer
2000;3:481–498.
26. Flack HD, Bernardinelli G. Reporting end evaluating absolute structure and
absolute configuration determination. J Appl Crystallogr 2000;33:1143–1148.
31. Levkin PA, Torbeev VY, Lenev DA, Kostyanovsky RG. Homo- and het-
erochirality in crystals. Top Stereochem 2006;25:81–134.
27. Gernez. Se´paration des tartrates gauches et des tartrates droits a`
l’aide des solutions sursature´es. Compt Rend 1866;63: 843.
Chirality DOI 10.1002/chir